Discriminate of microcalcification in breast tissue with digital mammography: the presentation a new method of processing

Introduction . Techniques developed in compute r and automated pattern recognition can be applied to assist radiologists in reading mammograms. With the int roduction of direct digital mammography, this will become a feasible approac h. A radiologist in breast cancer screening can use findings of compute r as a second opinion, or as a pointer to suspicious regions. This may increase the sensitivity and specificity of screening programs and it may avoid the need for double reading . Materials and Methods. A program for de tecting microcalcification clusters has performed, which discriminates clusters from normal breast tissue. First, we eliminated amorphous "clouds" or "blobs" in mammograms produced by normal paranchymal tissue of varying density using local average subtraction. Then we identified and removed the normal breast tissue. We have applied opening morphology operator and then, closing morphology in residual image. Any microcalcification th at may exist in mammogram is therefor e enhanced in the re sidu al image, which makes the decision regarding the microcalcification of mammogram be much easier. The digital images were presented to radiologists before appling algorithm and after it. Results. Res ults of study show th at sensitivity of this method in diagnosis is 100% against conven tional mammogram (85.4%). Conclusion. Our results have demonstr ated th at this algorithm can be an effective aid to radiologists in the detection of a range of types of microc alcification in mammograms in an environment that is similar to routine clinical screening.